PRIMED · DEVELOPER PLATFORM
Developer
Documentation.
Developer documentation, API references, and Model Context Protocol (MCP) server specifications for Primed athletic coaching intelligence.
FEATURED SPECIFICATION
Primed Model Context Protocol (MCP) Server
The Primed MCP Connector provides an authenticated, read-only interface allowing AI assistants (such as Meta Muse, Claude, and OpenAI ChatGPT) to access an athlete’s canonical coaching decisions, planned training schedules, athletic recovery states, and completed workouts.
CANONICAL SERVICE ENDPOINTS
Endpoints & Discovery
CORE PRINCIPLES
Architectural Guarantees
Deterministic Authority
All coaching decisions originate from Primed’s deterministic engine (DailyCoachingDecisionService). External AI assistants consume and explain decisions; they never invent workouts or diverge from canonical records.
Zero Split-Brain Invariant
All surfaces (iOS, Mac, Web, MCP assistants) consume identical canonical records. Dual-writes, out-of-band mutations, and unversioned fallbacks are prohibited by design.
100% Read-Only Contract
The connector exposes zero mutation or deletion tools in V1. It cannot alter training calendars, modify athlete profiles, or execute financial transactions.
Non-PHI Privacy by Design
Exposes only high-level athletic training parameters. Collects zero Protected Health Information (PHI) under HIPAA and no clinical diagnostic data.
Client Configuration
Ready-to-use configuration examples for Claude Desktop, Cursor, and custom HTTP/SSE MCP client implementations with token management.
RFC 7807 Error Handling
Standardized JSON-RPC machine-readable canonical codes, actionable recovery steps, and strict rate limit enforcement.
DEVELOPER SUPPORT
Questions or Integration Inquiries?
Our engineering team supports developers building MCP clients and assistant integrations.
SECURITY & COMPLIANCE
Security Incident Reporting
Confirmed incidents involving athlete data are escalated within 48 hours per Meta Policy § 4.8.